Concordance between bulk and sc DE tables

[Nayrouz109]

Hello,

I have a question about measuring the concordance between bulk and sc DE tables. In your script, you used 6 different methods to generate DE tables when the input is bulk RNA-seq data. Which one is then used to measure the concordance with sc DE tables? or did you take the average of their concordance measure?

Thanks !!

[skinnider]

Yes, see the methods:

DE analysis of bulk RNA-seq datasets was performed with six methods (DESeq2-LRT, DESeq2-Wald, edgeR-LRT, edgeR-QLF, limma-trend, and limma- voom), except for the two pulmonary fibrosis datasets15; for these datasets, the raw bulk RNA-seq data from sorted cells could not be obtained, so only the results of the bulk DE analysis performed by the authors of the original publication were used. The AUCC and rank correlation were calculated for each bulk DE analysis method separately, and subsequently averaged over all six methods. DE analysis of normalized bulk proteomics data was performed using the moderated t-test implemented within limma, as in the original publication.